110 likes | 239 Views
Coherence of LLO and LHO Power Monitors LIGO-G010313-00-Z Sergey Klimenko August 15, 2001 Outline Parameters of PM signals Coherence Coherence time scale Results Conclusion. Approximately 10 hour of data, starting at GPS time 668212508 was used in this analysis.
E N D
Coherence of LLO and LHO Power Monitors LIGO-G010313-00-Z Sergey Klimenko August 15, 2001 Outline Parameters of PM signals Coherence Coherence time scale Results Conclusion
Approximately 10 hour of data, starting at GPS time 668212508 was used in this analysis. The parameters of 60Hz PM signals were measured using data segments of one second long. For each data segment , the average amplitude and phase Y were measured with LineMonitor The power frequency was estimated as a derivative of the phase Y(t). Parameters of 60Hz signal Amplitude variation ~10-3 L0:PEM-LVEA_V1 (black) H0:PEM-LVEA2_V1 (red)
S(t) = a(t)cos(Y(t)) for monochromatic signals Y(t) is a linear function of time LLO LHO Phase
During one second time interval the power frequency doesn’t change much. The Line Monitor measures average frequency, which is used as an estimate of instantaneous power frequency. The measured frequency may vary with time Frequency Power frequency for the L0:PEM-LVEA_V1 (black) and H0:PEM-LVEA2_V1 (red) channels
A sum of two harmonic oscillations sL(t) and sH(t) with the same frequency is also a harmonic oscillation the amplitude A is given by the average (over the time interval T) square amplitude is If the phase difference remains constant during time T, the signals sL(t) and sH(t) are coherent. If the phase difference changes randomly in time the sL(t) and sH(t) are not coherent and the interference term is zero. Coherence coefficient Coherence time scale Coherence
cos(Df) as a function of time Df distribution integrated over 10 hours Phase Difference more coherent less coherent
coherence of L0:PEM-LVEA_V1 and L0:PEM-EX_V1 Df rms ~ 1mrad Power coherence for the same site
Ts – time scale m =Ts/DT– number of samples DT – sampling interval Tc = DT = const for white noise Coherence Time Scale power monitor signals (LHO, LLO) L0:PEM-LVEA_V1 x L0:PEM-LVEA_V1 Tc Tc (mDT) Is there a coherence at time scale >60sec? • Ttot/Tc – effective number of samples
the coherence of signals sL(t) and sH(t+t), where t is a time delay between two signals. coherence coefficient significance level phase difference uniformity Coherence & Significance Level blk – power monitors (60Hz) red – magnetometers (60Hz) ~1000sec
One possible explanation of the observed coherence In ideal case the phase of each monitor is f(t)=w0t + const In real life: f(t)=w0t +j(t) +const, where j(t)is (hopefully) a random phase If j(t)=rcos(nt+q)+h(t) and n is the same for LLO & LHO, If qLand qH are constant, frequency n can be seen at Df Fourier spectra. Interpretation modulation @ ~0.5, 1., 2. mHz
60Hz power line is coherent at time scale <1min. The coherence time is ~42sec Long term (>1min) correlation between LHO and LLO 60 Hz power lines is observed Although, there is some indication of power correlation between the LHO and LLO sites for this particular interval of time, they may not be coherent in a longer run. To conclude if there are periods of time when the LLO-LHO coherence time is much longer then 1 minute, 24 hours of data for different days of week should be analyzed. Conclusion